Rangarajan Parthasarathy, CPIM | January/February 2013 | 23 | 1
A hybrid model for improved care
The US health care industry is one of the biggest in the world. According to the Bureau of Labor Statistics, health care is made up of 595,800 establishments, of which approximately 6,000 are hospitals. These hospitals employ 35 percent of all workers, and 72 percent of hospitals employ more than 1,000 people.
Hospitals are expected to provide excellent health care at a reasonable cost, which is quite a challenge considering the overall structure of the industry and the role of insurance companies in health care. Therefore, there is a pressing need to cut costs and improve operational efficiencies in hospitals. Realizing this, hospitals the world over—and particularly in the United States—have adopted modern techniques, including six sigma, lean operations, quality function deployment, and radio frequency identification (RFID). On the cost-cutting front, hospitals have started maximizing resource utilization and optimizing inventory control.
The EOQ model
Materials management and inventory control are functions as critical in hospitals as they are in any other business. The objective is to stock only as much material as needed to successfully conduct day-to-day business, because stocking excess material increases storage and upkeep costs. However, if there is a material shortfall, it can negatively affect the business and—in the case of hospitals—even cause loss of human life. Thus, there is a cost associated with not stocking enough materials. In addition, intangible expenses such as loss of customer goodwill and legal implications are significant.
Other relevant inventory management issues in hospitals include supplier lead time, the reorder quantity versus the ordering cost, and the amount of buffer stock versus the cost of stockout.
The economic order quantity (EOQ) model takes into account these issues and proposes an order quantity that optimizes them. EOQ is calculated using the formula:
Here, AU equals annual demand in units per year, CO equals fixed cost per unit replenishment or the cost per order, CU equals the unit variable cost of the item, and I equals the annual cost of carrying inventory or the cost of $1 of the item in inventory for one year.
In most situations, the EOQ model provides a reasonably optimized order quantity. However, the model is subject to several assumptions, all of which may not be applicable in the context of the operations of a hospital. Noteworthy among these are the assumptions of a deterministic demand rate and long planning horizons, both of which may not apply to many items used in patient treatment.
Other assumptions, such as ignoring the advantages resulting from joint replenishment reviews, may further weaken the outcome of the model in the context of hospital operations. Some hospitals operate under a centralized supply chain system, where purchasing and materials management for a group of hospitals that come under a common management are done centrally and in an aggregated manner. In such situations, individual units may have seasonal demand variations and may not follow the same pattern as the aggregated demand—therefore, the EOQ model may not be the best solution.
The goal of forecasting is to predict future demand based on past demand. Statistical forecasting involves selection of a demand model underlying the observed demand pattern and using the model to forecast future demand. As with any statistical analysis, the reliability of the forecast depends on the quantity of data available, as well as its reliability. The inherent assumption is that the underlying model and its parameter values will not change during the forecast period.
Some popular forecasting methods include the simple moving average, weighted moving average, simple exponential smoothing, trending, and damped trending. The fundamental demand model for each of these techniques is different. The accuracy of the forecast depends on the selection of an appropriate demand model, so using the services of a statistician or an engineer in method and model selection is highly recommended.
Using forecasting in hospital inventory management addresses some of the disadvantages of the EOQ model. For example, EOQ may prove to be a valid way to order bandages and sutures from a cost perspective, but not from a service perspective, given the seasonality variations—for example, during snowy winters, risk of head injury may be more prevalent than in summer. However, there is a certain risk in using forecasting for centralized supply chain situations because demand patterns of individual hospital locations and of individual departments within those locations may not be accurately reflected by the aggregate demand model. The solution in this case is to use forecasting for individual hospital units and employ appropriate demand models dictated by departmental demand fluctuations.
ABC analysis is a well-known selective inventory control method, whereby items in the inventory schedule are categorized by their financial impact and annual consumption. This method is similar to the Pareto analysis used in quality control and six sigma techniques. In this model, “A” describes the 20 percent of items that account for 70 percent of the annual consumption, “B” represents the 30 percent of items that account for 25 percent of annual consumption, and “C” is the remaining 50 percent of items that account for 5 percent of annual consumption. In the overall scheme of things, A-class items should receive the most attention from the materials management team.
Employing ABC analysis may help expedite inventory control in hospitals, as A items are monitored and controlled closely, followed by B items. Meanwhile, outsourcing with a view to cut costs is a good option for C items. In addition, quality should take priority over cost for A, and preferably B, items. Thus, outsourcing may not be the best option here.
Demand prediction becomes an important issue for the As, followed by Bs, because carrying more buffer stock often proves to be exorbitant and financially damaging. For C items, approximate demand prediction may suffice.
A weighted purchasing policy may be used in conjunction with ABC analysis in order to optimize inventory carrying costs. For example, A items may have a lower reorder point but a more frequent delivery than B items. Cs could have the highest reorder points with the least frequent delivery schedule. Ultimately, inventory carrying costs, ordering costs, reorder points, and reorder levels are optimized using this method.
Inventory turns refers to the measure of how many times inventory is used during the year. Mathematically, average days to use the inventory equals 365 days divided by inventory turns.
Existing literature suggests that the ideal inventory turns for a hospital should be greater than 15. An inventory turn amount of 10 is considered very low. A hybrid model of inventory control and a disciplined approach to inventory management thus are necessary for hospital inventory stores.
While the EOQ model, forecasting, and ABC analysis individually may not accomplish the desired goals of a hospital inventory management system, a hybrid system could be the answer. Following are steps to implement:
Meeting the objectives
- Automate the inventory system. Bar coding and RFID for inventory automation offer numerous advantages. They enable tighter control and long-term cost cutting, use of location numbers can ensure traceability and eliminate theft and pilferage, and RFID facilitates the identification and removal of malfunctioning or outdated equipment. In addition, automation paves the way for quick availability of critical stockkeeping units (SKUs) and assists with Sarbanes-Oxley compliance.
- Perform an ABC analysis. This should be completed with the involvement of the materials management team, doctors, and nurses (in order to understand the relative importance of various SKUs).
- Perform forecasting analyses. Once the ABC analysis is complete, a forecasting analysis should be performed for all SKUs that are involved directly with treatment and clinical functions. The results should be reviewed with the doctors and nurses in order to determine the SKUs critical to patient service and primary hospital functions. It may be appropriate to place such SKUs in a special class by themselves.
- Automate the EOQ analysis. Simple software should be used to automate the EOQ analysis of all SKUs in the hospital. The output of this software would consist of the EOQ for each SKU, along with the reorder point and order quantity that optimize both; ordering costs; and carrying costs.
- Apply hybrid model rules to SKUs. Once all of the previous steps are complete, review each SKU individually and establish its relative importance with regard to the others. This process could be called an item value analysis. Again, this exercise should involve the doctors and nurses. The EOQ-supplied solution for A items may either be accepted or adjusted based on forecast predictions and the opinions of the doctors and nurses. For B items, the EOQ-supplied solution may generally suffice. And for C items, the EOQ-supplied solution should be examined closely. For certain C items that belong in a special class as described previously, the EOQ solution may be too conservative and may need to be increased by a rule of thumb based on forecast predictions. In this case, it is foreseeable that all three methods will lead to a multiplication factor that dictates reorder point, order quantity, and buffer stock. Though this exercise is time consuming, it will provide realistic results that help hospitals achieve desirable inventory turns.
Hospitals must work to optimize inventory levels to control costs. Keeping in mind that the primary function of a hospital is to provide excellent medical care, it is not always possible to optimize all aspects of inventory management. However, using this hybrid model of inventory control may make it possible to achieve substantial cost reductions and maintain excellent care.
Rangarajan “Raj” Parthasarathy, CPIM, is a quality manager and management consultant based in Chicago, Illinois. He holds a master’s degree in industrial engineering and an MBA, and he is also a Six Sigma Black Belt. Parthasarathy may be contacted at email@example.com
To comment on this article, send a message to firstname.lastname@example.org.